In February, the co -founder of Opeli and former director of Tesla Ai Andrej Karpathy invented a phrase that quickly caused fascination, debate, and even a small cultural change in the world of software development: coding coding.
What began as a post “there is a new type of coding, which I call” climate coding “, in which you fully undergo vibraca, you accept interpretations and forget that the code exists at all”-has switched to a new way of thinking about human cooperation-ai in programming. Today programmers, technical influence and technical influence and Even companies like IBM will look closer In what this experimental coding style means the future of software creation.
VIBE coding describes an emerging practice in which developers use AI tools not only for help, but as a co -creator or, more provocative, as those driving the development process. Instead of meticulous planning and controlling each aspect of the program, Atmospheric codes interact with large language models (LLMS) Like Chatgpt, Claude or specialized code assistants such as Github Copilot and Cursor. They describe what they want, accept AI suggestions with minimal intervention and just see what will happen.
Carpathy's own process included errors messages in sending a copy to artificial intelligence, rarely questioning the results and trusting the system “fix your own mistakes”. This is a relaxed, improvisational approach that priority treats speed, spontaneity and intuition – almost the opposite of traditional software engineering.
Coding appeal consists in its simplicity and performance. Developers can focus on wider design concepts and user sensations, and not to be drawn to technical details. Atmospheric coding seems to be supported by two very different groups: experienced programmers who can catch and fix errors if necessary, and total beginners who want to introduce an idea without writing a lot (or any) code.
This practice has already proved to be worth prototyping small games, applications and websites. This drastically lowers the entry barrier, enabling the person there is no background in functional programming, if basic, software.
Experts warn that although the atmosphere coding can work in the case of a hobby or pet projects, it is not yet reliable or safe enough for serious systems. The code generated by AI may include errors, ineffective logic, and even dangerous gaps in security-projects requiring an experienced man to detect and solve.
For now, the consensus consists in the fact that the atmosphere coding shines in the early stages of experiments and ideas, but one should not rely on the bases of the Critical Code.
Will AI finally take over most of the software? Probably. But even as the automation progresses, many believe that human developers will continue to play a key role in debugging, architecture, optimization and integration of business logic. As Joshua Noble put it, the technical strategist at IBM: “Fighting software engineering will always require man at some point.”
Because LLM is still improving, and AI coding assistants gain real -time integration with programming environments and cloud platforms, the nature of software creation may change permanently. The line between Koder and Koder is already blurred.
Coding atmosphere could start as a meme, but it quickly becomes a mirror reflecting our changing relationship with AI. He is chaotic, imperfect, sometimes stupid – and very human. Maybe that's it.
